Somewhere right now, an engineer at a streaming service is about to push a build. A vetted tester in Manila is about to find the bug that would have ruined her launch day. The tester is logged into Testlio. The engineer has probably never heard the name. That is, more or less, the entire business.
Testlio sits in a strange spot in software. It is not glamorous. It does not promise to disrupt anything. It does the quiet, unfashionable work of making sure the apps on your phone do not embarrass the companies that built them. And for the last twelve years, it has been doing that work for an unreasonable share of the brands you use every week.
Crowdtesting was broken. It paid testers per bug, then wondered why testers stopped caring.
In the early 2010s, crowdsourced testing existed, but it was a race to the bottom. Platforms paid pennies for each bug found, which incentivized noise: shallow reports, duplicate filings, the same five testers gaming the same five products. Quality was treated like piecework in a sweatshop. Predictably, the work felt like piecework in a sweatshop.
One of the testers grinding through that system was a 23-year-old Estonian named Kristel Kruustük. She was good at the job. She kept winning awards. She also kept feeling, correctly, that the system rewarded the wrong things.
Win a hackathon. Quit your job. Build the version of crowdtesting that doesn't treat humans like a search algorithm.
In 2012 Kristel and her co-founder (and now husband) Marko Kruustük entered AngelHack in London with an idea: vet the testers, train them, manage them, and pay them like professionals. They won. They took the prize money, moved to Austin for Techstars, and quietly began assembling the operating system for a different kind of QA company.
The bet was unsexy and counter-cyclical. While the rest of the industry was selling "automation will replace humans," Testlio was arguing the opposite: humans matter more, but only the right ones, in the right structure, with the right tools behind them. They called it fused testing long before fused became a buzzword.
Three things to know about the founders
1. Kristel made Inc.'s 2022 Female Founders 100 list. 2. The Kruustüks are married - the company and the marriage have outlived several of their early investors' funds. 3. Marko stays mostly out of press releases. He is, by all accounts, fine with this.
One platform. A global crowd. A lion logo.
What Testlio actually sells is hard to put on a billboard, which is part of why it doesn't try. There is a software platform that orchestrates tests, manages testers, files bugs, integrates with Jira and TestRail and BrowserStack, and produces dashboards that engineering leaders can read without crying. There is a managed services layer that staffs each project with vetted humans across 150+ countries. And, since 2025, there is a layer of AI - the LeoAI Engine, LeoMatch, and LeoInsights - that quietly does the matching and summarizing that used to require an analyst with a spreadsheet and a sigh.
The taxonomy is dull and the work is hard: functional testing, regression, exploratory, localization, usability, payments, video, livestream, accessibility. Test the app in Brazilian Portuguese on a three-year-old Android in a coffee shop with bad Wi-Fi. Test the streaming service on the smart TV nobody remembers shipping. Test the payment flow at 2am when your tester in Lagos can actually reproduce the failure mode.
Testlio, in twelve mostly-on-purpose years
- 2012Founded after winning AngelHack in London. The prize check becomes the seed capital.
- 2013Joins Techstars Austin. The Estonians arrive in Texas. The Texans buy them barbecue.
- 2014Seed round closes; Altos Ventures takes an early bet.
- 2018$6.6M Series A from Altos and Vertex Ventures US.
- 2021$12M Series B at a reported $100M valuation. Already profitable - which by 2021 standards counted as a personality trait.
- 2022Kristel Kruustük named to Inc.'s Female Founders 100.
- 2025Ships LeoAI Engine, LeoMatch, LeoInsights. Launches end-to-end AI testing service.
- 2026Summer Weisberg appointed CEO. Company posts a record year.
The customer list is the marketing.
Companies do not hand QA over to a vendor lightly. A bad test partner ships bad software, and bad software ships angry users. So the more interesting Testlio fact is not the funding total or the headcount - it is who keeps signing the renewal.
Where the money came from
Sources: Crunchbase, TechCrunch, Testlio press releases. Reported Series B valuation: ~$100M.
Make software people can actually trust.
You will not find Testlio writing manifestos about "the future of work" or "democratizing" anything. The pitch is narrower and more honest: the more software a company ships, the more ways it can fail in front of the people who pay for it. Testlio exists to find those failure modes first.
What Testlio is not
It is not a freelancer marketplace. It is not a no-code automation tool. It is not an offshore outsourcer in disguise. It is, if you must categorize it, a managed services firm with a real software product underneath - the kind of business public markets don't quite know what to do with, which is partly why it has remained quietly private.
The customer base reflects the pitch. Streaming services who cannot afford a Christmas Eve outage. Payments companies who cannot afford a stuck transaction. Healthcare platforms who cannot afford to mistranslate a dosage. Sports leagues who cannot afford a broken stream during a playoff game. These are not companies looking for cheap testing. They are looking for testing that works at 2am.
AI is shipping faster than anyone can test it. That is the next decade for Testlio.
Every company is now an AI company, which means every company is now shipping software that can be wrong in unfamiliar ways. Models hallucinate. Agents misroute. Chatbots invent policy. The traditional QA stack - written for deterministic software - was not built to catch any of this.
Testlio published its own data in late 2025: 82% of AI bugs its testers found were misinformation and accuracy failures, not crashes. That is a category of defect software companies have, historically, simply not had to worry about. The implication is straightforward and slightly inconvenient: somebody, somewhere, needs to verify that the AI in your product is telling the truth before it tells it to your customers.
Under new CEO Summer Weisberg, Testlio's bet is that this work scales. Not as a side feature of automation tooling, but as a full managed service - humans evaluating models, structured by software, summarized by AI, sold to the companies racing to ship features they barely understand themselves. It is, in a way, the same bet Kristel made in 2012. The technology has changed. The thesis - humans matter, but only with structure behind them - has not.
Back to the engineer about to push the build.
She still does not know Testlio's name. The launch goes out at 9am her time. Nothing catches fire. The release notes get a small round of internal applause. Someone in product brings donuts.
Somewhere in Manila, the tester who found the bug logs off, marks the task complete, and gets paid - not per bug, not as piecework, but as a professional whose job is to care about software for a living. That is the company Testlio set out to build. Twelve years in, it is mostly the company it has become.
The best QA company in the world is the one you never have to think about. Testlio appears to be making peace with that.